Non-invasive classification and imaging of septal and paraseptal ventricular preexcitation sites by contactless multichannel magnetocardiography

Non-invasive classification of ventricular preexcitation (VPx) with ECG algorithms is considered adequate to plan ablation approaches, with reported accuracy ranging between 80 and 92%. However during ablation, especially of septal (S) and paraseptal (PS) accessory pathways (APs), unexpected difficulties might arise, due to complex or multiple preexcitation. Therefore imaging techniques (BSPM, Echocardiography, SPECT) have been used to minimize the risk of failure. Contactless magnetocardiographic mapping (MCG) has been proposed, since 1984, as an alternative method for non-invasive three-dimensional (3D) localization of arrhythmogenic substrates and in particular of VPx. The aim of this study was to compare the accuracy of MCG classification of S and PS VP with that obtained with 7 most used ECG algorithms. Method: 42 patients, with ECG consistent with septal VPx, were studied with 36-channels DC SQUID MMCG system (sensitivity is 20 fT/Hz\ubd, at 1 Hz) recording the z component of magnetic cardiac field from a 36-point grid (20 x 20 cm). The total time for MCG cardiac mapping, automatic source localization and 3D imaging was typically 90 seconds. MCG localization of VPx was obtained with the Equivalent Current Dipole (ECD), the Effective Magnetic Dipole (EMD) and the Currents Reconstruction (CR) models. 3D multimodal imaging consisted of MCG, MRI and fluoroscopy data fusion. Successful catheter ablation, when clinically required (12 cases), was the goldstandard for validation. Results: MCG classification of VPx was certain in 38/42 pts (90.5%) and was confirmed in all 12 pts undergoing successful ablation. ECG classification was certain in 29/42 (69%) pts, but uncertain in 13/42 (31%). MCG provided a clear-cut localization in 7/11 pts with uncertain ECG and demonstrated complex activation patters during the delta wave in the remaining 4, consistent with possible branching of septal pathways, unpredictable on the basis of ECG. Conclusion: As compared to ECG, MCG classification of S and PS VPx is more accurate because it adds: 1) information on the origin and the direction of the preexcitation wavefront across the septal area; 2) 3D electroanatomical imaging of the VPx site into a 3D model of the patient's heart reconstructed from MRI, with easy differentiation among superior (anterior), inferior (posterior) and midseptal APs. In the EP laboratory, the 3D heart model with MCG localization can be interactively rotated to match with 2D fluoroscopic projections, thus facilitating anatomically appropriate mapping and positioning of the ablation catheter, without other navigation systems